The epidemiology of preterm birth.

Secular trends in the prevalence of preterm birth and international comparisons of the rates of preterm birth are difficult to interpret because of differences, both formal and informal, in the registration of extremely preterm births. Accurate estimation of gestational age is another problem in the measurement of preterm birth. Preterm birth is heterogeneous in several ways. It is heterogeneous in terms of the extent to which the birth is preterm (20-27 weeks, 28-31 weeks or 32-36 weeks of gestation); in whether the birth was elective or spontaneous; and among spontaneous idiopathic preterm births, in whether there was preterm labour or premature rupture of the membranes. Case-control study designs taking account of these subgroups have been a recent feature of epidemiologic approaches. The classic social associations of preterm birth--low socioeconomic status, extremes of maternal age, primiparity, being unmarried--apply to extremely preterm and moderately preterm births as well as to the mildly preterm group. The strength of these associations is small compared with factors in the prior reproductive history and with medical and obstetric complications of the current pregnancy. Recent epidemiological research activities have focused on the ways in which risk factors such as physical workload, drugs and alcohol, lack of social support and infection might be mediating factors between sociodemographic status and preterm birth. As Eastman (1947) pointed out almost 50 years ago, 'only when the factors causing prematurity are clearly understood can any intelligent attempt at prevention be made'.